Lifetime reduction when switching on/off often

[E. Jongerius]

Hi guys,

Officially I am an electronics engineer and for one of my customers I am developing an electronics controller which can held a (original) domestic refrigerator within +- 0.1C bounds of its setting value.

The best way I found to achieve this is to implement a PI(D) controller with a cycle time of about 1 minute, and let the duty/cycle control the motor (eg 50% D/C means 30sec on, 30sec off). The motor is an RSIR type compressor, and I have removed the external starter PTC (it was too slow) and built an external circuit to start the motor. I have not yet experienced a blocked roto due to the pressure buildup after the compressor, so this works great.

However, the compressor is switched on very often relatively to its original design. Does this shorten its lifetime significantly or is this not a real issue?

Thank you,

[The Viking]

As the system wouldn't equalise the pressures in the cycling times you describe, chances are that it wouldn't start for a second cycle at all.
If it works, the compressor would get loads of high current starts but not the cooling from the refrigerant so it's likely to burn windings at a rapid rate. In 30s it is unlikely that you even get a proper refrigeration cycle going...

In climate chambers the standard is to keep the compressor running and offload the circuit with bypasses.

[E. Jongerius]

That should be a very good solution, however these refrigerators may cost very little. It should be a very big issue for my customer if he has to change the plumbing, as it implies lots of extra work on each machine. At this moment the only thing he has to change are the thermostat (remove), add the electronics and do some wiring.

I did a few tests with restarting the compressor when it had just stopped 2 seconds (thus not equalising the pressures) and that does really work well. Test is running now for about 15 hours, switching on/off every minute and I haven't seen a problem yet.

So I am able to restart it, but are there anymore downsides doing this?

[Brian_UK]

I would suggest that you monitor closely the compressor shell temperature and the motor running amps.

This will show you the state of the compressor and indicate when it is about to seize or burn out.

[E. Jongerius]

Ok, that I will do. Are there any guidelines for this?

Motor amps (main winding) are very constant when running. I do not have logged the shell temp, but it seems to vary a bit. I think maximum measured is about 50C.

it is a Chinese made 1/10hp 230V~ CSIR compressor, 5,3ccm displacement.

[desA]

As the system wouldn't equalise the pressures in the cycling times you describe, chances are that it wouldn't start for a second cycle at all.
If it works, the compressor would get loads of high current starts but not the cooling from the refrigerant so it's likely to burn windings at a rapid rate. In 30s it is unlikely that you even get a proper refrigeration cycle going...

This is very, very important. It takes time for a refrigeration cycle to stabilise - it's a 'slushy' technology...

[mad fridgie]

I do not see a direct problem with refrigerant pressure equalization when you have a compressor with a dedicated start winding and control (not to be confused with a PSC start as used in AC systems)
Look at all the pump down systems without start unloading.
I do see 2 problem increase were and tare on the mechanical componets, due to increases in starting thrust(s) but main excessive heat build up in the winding (stator).
At the point of start up you basically have a dead short (resistive) only when rotation occurs does it become inductive. What is unknown; is there enough internal mass to disiapate this heat, without the need for refrigerant mass flow. I think measuring the shell in this case is not suitable (motor/comp is hung within the shell and is not in direct contact, I believe the shell is also on the low side, perhaps measuring the discharge temp may be a better option. If the motor is indeed is very hot, then the actual vapor entering the internal suction valve (head) will be highly superheated, over what enters the pipe suction pipe on the compressor.
I would not produce and market a mass product that has only been tested for 15hrs. (normal run time of a fidge is well in excess of 65000hrs)

[hookster]

Interesting concept to require close control on a refrigerator that has a door opening of 1/4 of cabinet wall structure with no secondary refrigerant system to maintain close control upon failure.

The short cycling of compressor will affect oil returns back to the compressor but as you say it is not cost effective to redesign refrigerant layout.
So a second backup refrigerator on standby waiting to go will be required.

[HallsEngineer]

For gods sake use a small VFD and reduce the speed of your machine linked to suction pressure. ON/OFF NOT GOOD!!!

[Frikkie]

Most motors are designed to make a certain number of starts per hour maximum. To many starts makes too much heat this would be even worse if starting is under load which makes it slow starting.

Too much heat makes poor winding insulation and might cause oil breakdown. All this means premature failure.

I have seen small oil cooler before on domestic fridges, maybe one of these could help. It might also be better to monitor oil temperature as an indicator of system stress rather than dome/shell temperature. There might be a correlation between the two but is difficult to know what it is.

[lowcool]

i assume it has to reach set point,then cycle as you say.as all is listed above i would be concerned.maybe fit a start cap when in close cycling mode,this will cause extra heat in the windings but reduce contact wear and tear if a relay is switching comp relay.definitely no cap in pull down mode and maybe have extended run period here and there for a cooling effect,oil return.i cannot see how refrigerant will stabilise with continuos off continuos on for that time period.to me everything may seem alright if the door is closed all the time and no load is placed on the system.to me normal off cycle creates similar power consumption if not less

i think your mate would be better off reinventing the esky,better chance of getting a cold beer out of it all day every day
cheers